How the #1 Women's Health App Keeps Performance Fast and Errors Down for Millions Globally

TL;DR Flo Health, makers of the world’s leading women’s health app, uses Sentry to ship fast and fix faster. With full-stack visibility across mobile and backend, Flo catches issues before they hit production, tracks performance in real time, and keeps millions of global users engaged—all without slowing down their 50+ deploy/day cadence.

Results:

• Increased request velocity for key transactions by 50%
• Maintained 99.9%+ crash-free rate across millions of users
• Enabled confident mobile rollouts and fast rollback strategies
• Supported ~50 backend deploys/day without sacrificing quality or visibility

If our app is slow, users won’t wait—they’ll just leave. Using Sentry to debug and monitor performance is a key factor to our success in keeping users in our app and engaged. — Vaidas Zlotkus, Director of Engineering

From Backend Deploys to Edge Latency: Debugging Responsiveness for 100M+ UsersFrom Backend Deploys to Edge Latency: Debugging Responsiveness for 100M+ Users

Flo is the most widely used women’s health app globally, offering AI-powered insights and medically reviewed content in multiple languages. Their infrastructure powers a backend-driven UI, supports tens of thousands of content articles, and delivers frequent backend updates—up to 50 times per day. Ensuring application speed and stability at this scale is not just a technical goal, but a business imperative.

With users around the globe, Flo must meet high standards of responsiveness, data protection and reliability regardless of region or device. This requires robust observability tools that detect issues, apply robust privacy and data minimisation protections, and empower teams to act on them—fast. Sentry provides that full-stack visibility.

How Flo Uses Sentry to Deliver a Fast and Stable AppHow Flo Uses Sentry to Deliver a Fast and Stable App

1. Proactive Release Monitoring & Crash Management1. Proactive Release Monitoring & Crash Management

Flo’s engineers use Sentry as their first line of defense during mobile rollouts. With staggered deployments and varied user adoption rates across versions, it’s critical to detect regressions early—before they escalate.

When we do a 5% mobile rollout, checking Sentry is part of our release checklist, said Vaidas. We’ve caught major issues this way and stopped bad releases before they could impact our user base.

Engineers monitor crash trends by version, not just overall rates. A “sheriff rotation” ensures that dedicated team members keep an eye on stability metrics during rollouts, enabling timely interventions when anomalies occur.

2. Debugging region-specific regressions2. Debugging region-specific regressions

At global scale, performance is a moving target. New devices, flaky networks, edge compute, and a dozen other variables—all conspiring to make apps feel slower than they should. Flo understands that delivering a fast experience means looking beyond backend metrics and into what users actually feel.

To get there, Flo uses Tracing and augments traces with custom RUM implementation data to measure the full journey: from backend response to UI rendering. This includes all the stuff that usually hides in the “it’s probably fine” pile—like asset load times, cold starts, and performance issues in specific locales or languages.

Sentry helps us bridge the gap between what the server sees and what the user feels, said Vaidas Zlotkus, Director of Engineering. By combining tracing and session data, we can pinpoint bottlenecks and fix them—whether they’re on the backend, at the edge, or on the device.

The team monitors all the usual suspects—p95 durations, apdex scores, 1s timeout rates—but tracing gives them the extra context to understand why and where things are slow, not just that they are. That clarity has led to measurable impact:

• Backend p95 latency dropped by up to 200ms
• Client-side load times now consistently under 350ms (p95)
• Apdex firmly in “excellent/good” territory
• Timeouts down to 1% or less on key feature toggle flows

These gains didn’t happen by accident. They happened because Flo builds with performance in mind, iterates quickly, and doesn’t settle for “fast enough.”

Latency is key, said Vaidas. If our application is slow, users will ditch it.

Speed is a key driver of user engagement, and Flo approaches performance like a product feature—not a one-time fix. Using Sentry, the team tracks latency percentiles (P50, P95, P99), cold vs. warm app starts, and other real-world indicators that affect how users experience the app day-to-day.

When experimenting with edge compute (via Cloudflare Workers) to accelerate web onboarding flows, Sentry gave them region-specific performance insights. Dashboards segmented by country helped guide decisions like which users should hit the edge vs. go direct to AWS—something native observability tools didn’t always provide.

Sentry helped us evaluate latency tradeoffs when experimenting with edge computing, which doesn’t always have built-in observability, added Roman Bugaev, CTO of Flo Health.

3. Debugging Critical Backend Services3. Debugging Critical Backend Services

Flo’s core services are written in Python, and during outages or unexpected backend behavior, engineers turn to Sentry first.

During any major incident, Sentry is the first thing we look at to understand what’s broken and where, Vaidas confirmed.

Stack traces, error frequency, and environment tagging make Sentry indispensable in real-time triage and long-term root cause analysis.

4. Uncovering Edge Cases & Content Issues4. Uncovering Edge Cases & Content Issues

With over 10,000 content articles delivered via Contentful and localized in numerous languages, content-specific bugs can be hard to detect—especially when they only affect edge cases.

We’ve built dashboards in Sentry to track content-related issues, Roman explained. For instance, if a story fails to render or a video doesn’t load in Portuguese, we catch that through custom event tagging in Sentry.

This has helped the team identify errors tied to specific locales, content types, or edge formatting rules that could otherwise go unnoticed.

More engineering focus, not monitoring overheadMore engineering focus, not monitoring overhead

By embedding Sentry into the daily workflows of both backend and mobile engineers, Flo ensures that visibility into errors and performance is never an afterthought. It’s integral to how they operate.

Sentry enables Flo’s engineers to:

• Deploy frequently and safely
• Detect regressions before users are impacted
• Monitor app performance across regions and platforms
• Investigate and fix issues across backend and mobile systems
• Track content-specific and locale-specific errors that impact UX

Sentry helps us focus on what matters: building a better health app. We don’t want to spend our time maintaining observability tools—we want to use them to move fast and improve the experience for our users, concluded Roman.